Munster: Tesla's autonomy future is in hardware

Tesla's autonomous vehicle strategy is increasingly centered on computing hardware advancements, with Deepwater Asset Management's Gene Munster detailing the progression from HW3 to HW4 and the critical upcoming HW5 for achieving Level 4/5 autonomy and scaling its robotaxi fleet. This hardware-centric approach may incur a manageable $40-80 million in potential FSD refunds for HW3 customers, while underscoring the ongoing demand for continuous hardware upgrades to meet escalating autonomous driving computing requirements.

Analysis

Tesla's autonomous vehicle strategy is undergoing a significant pivot, shifting focus from battery range to the computational power of its in-car hardware as the primary value driver. According to analysis from Deepwater Asset Management, this transition is defined by distinct hardware generations: HW3, currently in approximately 1.4 million U.S. vehicles, is considered insufficient for advanced autonomy, while the newer HW4, found in 1.2 million vehicles, offers 3-8 times the computing power and underpins the current expansion of Tesla's Austin robotaxi fleet from 15 to 35 vehicles. The forward-looking strategy hinges on the upcoming HW5, expected early next year with a projected tenfold increase in processing power over HW4, which is seen as the necessary platform to achieve Level 4 or Level 5 autonomy and scale a competitive robotaxi service against incumbents like Uber. A potential financial headwind from this hardware-centric approach is a $40-$80 million liability for refunding Full Self-Driving (FSD) purchases to HW3 owners, though this is considered a manageable expense against the company's $37 billion cash and investment position. The outlook for deploying over one thousand robotaxis by year-end is viewed as realistic, signaling confidence in the current HW4-based capabilities as a bridge to the more advanced HW5.

Key Decisions for Investors

• Investors should monitor the development and rollout timeline of HW5, as its launch and subsequent performance are presented as the primary catalysts for achieving Level 4/5 autonomy and unlocking the robotaxi business valuation.
• The potential FSD refund liability of $40-$80 million for HW3 vehicles should be considered a minor, one-time cost of transitioning the fleet to superior hardware, not a material threat to Tesla's strong liquidity position.
• Track the expansion of the robotaxi fleet towards the year-end goal of over a thousand vehicles, as meeting this target would validate the current HW4-based strategy and demonstrate tangible progress in competing with established ride-sharing networks.